Content
Runoff and Stream Water Chemistry Responses to Simulated Emerald Ash Borer Invasion in Black Ash Wetlands in Northern Michigan
Joseph Shannon1, Kathryn Hofmeister1, Matthew Van Grinsven1, 2, Randall Kolka3, Fengjing Liu1
1 College of Forest Resources and Environmental Science, Michigan Technological University; 2 Department of Earth, Environmental, and Geographical Sciences, Northern Michigan University; 3 US Forest Service, Northern Research Station
github.com/jpshanno/agu_2019
jpshanno@mtu.edu
@jpshanno
Major Finding
Stream water dissolved organic carbon concentrations increased following disturbance. Changes were driven by increased wetland water levels and a more continuous surface water – stream water connection.
Basic Sections
Introduction
Black ash is a dominant canopy species in northern forested wetlands that is threatened by Emerald Ash Borer (EAB), which is continuing to spread into areas where black ash is prominent and loss of ash has been shown to affect wetland water levels. No previous work has examined water quality changes following simulation or how those changes are telegraphed downstream.
Methods
A paired watershed approach has been implemented on the Ottawa National Forest in western Michigan (Control: 1.1 ha, 503 m a.s.l.; Treatment: 0.8 ha, 476 m a.s.l.). Streamflow was monitored in a 6” Parshall flume with a Solinst Levellogger Edge and dissolved organic carbon (DOC) and total dissolved nitrogen (not shown) were collected by grab and automated sampling and processed on a Shimadzu Total Organic Analyzer. After 2 years of calibration all ash ≥ 1” dbh were cut and left on site in one watershed in March 2015. Changes in wetland water levels (Figure 1) were tested using Kruskall-Wallis Rank Sum tests. Hydrograph separation (Figure 2) was performed using three approaches to allow for sepearate measures of baseflow and quickflow (Lyne and Hollick (1979); Chapman and Maxwell (1996); and Eckhart (2005)). Pre-treatment models for stream water DOC concentration between sites (Figure 3), and for stream water DOC concentration by wetland water source (Figure 4 & Table 1) were fit using quantile regression models with \(\tau = 0.5\). Quantile regression models avoid undue outlier influence on model fit and account for unmodeled environmental drivers (Cade and Noon 2003).
Summary Results
During the post-treatment period wetland water levels in the treatment watershed increased relative to pre-treatment conditions. During the same period the percent of water yield as baseflow increased relative to the control watershed. Increased wetland water levels likely led to a more continuous surface water – stream water connection, resulting in higher baseflow. This is supported by the wetland surface water model at the treatment site, which was the best performing post-treatment predictor of stream DOC. Over that same period in the control watershed baseflow decreased and wetland soil water was the best predictor of stream water DOC.
Future Work
Future work will incorporate seasonality as these systems are heavily influenced by snowfall, and analyze anion (Cl-, SO42-) concentrations and flourescence excitation emission matrices to improve source water separation and understand biogeochemical processing.
Cade, Brian S., and Barry R. Noon. 2003. “A Gentle Introduction to Quantile Regression for Ecologists.” Frontiers in Ecology and the Environment 1 (8): 412–20.
Chapman, T.G., and A.I. Maxwell. 1996. “Baseflow Separation - Comparison of Numerical Methods with Tracer Experiments.” In Hydrology and Water Resources Symposium, 539–45. Hobart, Australia.
Eckhart, K. 2005. “How to Construct Recursive Digital Filters for Baseflow Separation.” Hydrological Processes 19: 507–15.
Lyne, V, and M Hollick. 1979. “Stochastic Time-Variable Rainfall-Runoff Modelling.” In Institute of Engineers Australia National Conference.
Thorne, W. Brent. 2019. Posterdown: An R Package Built to Generate Reproducible Conference Posters for the Academic and Professional World Where Powerpoint and Pages Just Won’t Cut It.